HappyDB is a corpus of 100,535 crowd-sourced happy moments. The goal of this project is to look deeper to compare the causes between the short term happy moment and long term happy moment. I apply natural language processing, text mining and data manipulation to derive interesting findings in the collection of happy moments.
## Warning: package 'bindrcpp' was built under R version 3.4.4
Firstly, I check the number of last 24 hours happy moment is 49831 which is very similar with the number of last 3 months happy moment: 50704.
24 hours country frequency
##
## USA IND CAN GBR VEN PHL MEX AUS NGA
## 5594 694 42 35 34 34 17 14 8 8
3 months country frequency
##
## USA IND CAN VEN GBR PHL AUS BRA MEX
## 6118 708 41 37 37 25 24 12 11 11
24 hours marital frequency
##
## single married divorced separated widowed
## 3461 2718 333 49 42 31
3 months marital frequency
##
## single married divorced separated widowed
## 3845 2878 352 70 44 26
24 hours gender frequency
##
## m f o
## 3347 3230 42 15
3 months gender frequency
##
## m f o
## 3654 3511 34 16
24 hours parenthood frequency
##
## n y
## 3926 2691 17
3 months parenthood frequency
##
## n y
## 4355 2847 13
24 hours age frequency
##
## 23 21 25 29 19 33 27 35 37 16
## 224 220 215 205 192 190 183 182 179 174
3 months age frequency
##
## 23 25 21 33 29 27 19 35 37 22
## 244 236 232 227 219 216 203 188 181 177
From above frequency tables, we can see that the most frequent class in each category of workers are very similar and share similar number. So it can be concluded that comparison on short term happy moment and long term happy moment is not affected by country, marital, gender, parenthood and age. So next, we can begin to compare.
The Happiness category label predicted by the author (Please see the reference for details) which is 7 categories: achievement, affection, bonding, enjoy_the_moment, exercise, leisure, nature. Below is the percentage of each category appears in the collection of the last 24 hours happy moment as well as in the collection of the last 3 months happy moment.
24 hours happy moment Happiness category label
##
## achievement affection bonding enjoy_the_moment
## 30.980715 32.818928 10.477414 13.341093
## exercise leisure nature
## 1.531175 8.695390 2.155285
3 months happy moment Happiness category label
##
## achievement affection bonding enjoy_the_moment
## 36.5947460 35.1333228 10.8591038 8.8671505
## exercise leisure nature
## 0.8658094 6.1632218 1.5166456
From above group barplot and pie charts, we can see that when people talking about their happy moment in last 3 months, they are more focus on achievement, affection and bonding. While, when people talking about their happy moment in last 24 hours, they are more focus on enjoy the moment, exercise, leisure and nature.
24 hours most 20 frequent words
##
## friend watched night played home dinner morning feel
## 5035 2694 2374 2236 2234 2208 2065 2010
## family enjoyed found son game finally daughter nice
## 1932 1901 1751 1735 1694 1652 1649 1647
## favorite hours life love
## 1617 1568 1440 1403
3 months most 20 frequent words
##
## friend family finally job home found feel birthday
## 5857 2760 2270 2162 1977 1969 1936 1934
## son play daughter bought watching enjoyed event life
## 1898 1822 1759 1747 1691 1663 1557 1555
## love received game started
## 1553 1514 1510 1489
24 hours Word Cloud
3 months Word Cloud
Friend, family, son, daughter are always the most common words in the happy moment both in last 24 hours and last 3 months. So obviously the key to be happy is spending more time with friends and family.
In the last 24 hours dataset, watched, game appear more frequently. Which means game, tv show, film are really a good way to relax people and make them feel happy immediately in daily life. Additionally night, morning are also very common in the last 24 hours. So a good begining and ending of the day can make someone’s day. Moreover dinner is the 6th most frequent words in the last 24 hours dataset, which we can see that dinner is the most important meal in a day and food can make people feel good quickly.
In contrast, event, birthday appear more frequently in the last 3 months. It is resonable because people easily think over and remember some great event and something can change their life when they are asked about the happy moment in a longer period. People needs holiday and event to refesh themselves and get reunion with their family and friends to make them feel great. The other common words are job, bought. It’s also imaginable because job is really important in the life and shopping is usually a way to decompress. And I feel a little surprised to see bought is in the common words of the last 3 month instead of the last 24 hours. This shows shopping is a good way to be relaxed and help people be happy in a long term.
Akari Asai, Sara Evensen, Behzad Golshan, Alon Halevy, Vivian Li, Andrei Lopatenko, Daniela Stepanov, Yoshihiko Suhara, Wang-Chiew Tan, Yinzhan Xu, ``HappyDB: A Corpus of 100,000 Crowdsourced Happy Moments’’, LREC ’18, May 2018. (to appear)